We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations

DBLP - CS Bibliography


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computer Vision and Pattern Recognition

Title: Greedy Gradient Ensemble for Robust Visual Question Answering

Abstract: Language bias is a critical issue in Visual Question Answering (VQA), where models often exploit dataset biases for the final decision without considering the image information. As a result, they suffer from performance drop on out-of-distribution data and inadequate visual explanation. Based on experimental analysis for existing robust VQA methods, we stress the language bias in VQA that comes from two aspects, i.e., distribution bias and shortcut bias. We further propose a new de-bias framework, Greedy Gradient Ensemble (GGE), which combines multiple biased models for unbiased base model learning. With the greedy strategy, GGE forces the biased models to over-fit the biased data distribution in priority, thus makes the base model pay more attention to examples that are hard to solve by biased models. The experiments demonstrate that our method makes better use of visual information and achieves state-of-the-art performance on diagnosing dataset VQA-CP without using extra annotations.
Comments: Accepted by ICCV 2021. Code: this https URL
Subjects: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL); Machine Learning (cs.LG)
Cite as: arXiv:2107.12651 [cs.CV]
  (or arXiv:2107.12651v4 [cs.CV] for this version)

Submission history

From: Xinzhe Han [view email]
[v1] Tue, 27 Jul 2021 08:02:49 GMT (2174kb,D)
[v2] Tue, 3 Aug 2021 01:26:10 GMT (2176kb,D)
[v3] Mon, 9 Aug 2021 13:36:51 GMT (2176kb,D)
[v4] Mon, 23 Aug 2021 12:11:39 GMT (2177kb,D)

Link back to: arXiv, form interface, contact.